基础信息
项目名称:hpcaitech/ColossalAI
项目徽章:
仓库地址:https://github.com/pterodactyl/panel
检测报告地址:https://www.murphysec.com/console/report/1718667467391025152/1718667510734962688
此报告由Murphysec提供
漏洞列表
漏洞名称 | 漏洞类型 | MPS编号 | CVE编号 | 漏洞等级 |
---|---|---|---|---|
低危 | ||||
NumPy 代码问题漏洞 | 空指针取消引用 | MPS-2021-32278 | CVE-2021-41495 | 中危 |
Python 安全漏洞 | ReDoS | MPS-2022-57238 | CVE-2022-40897 | 中危 |
PyTorch 命令注入漏洞 | 代码注入 | MPS-2022-65270 | CVE-2022-45907 | 严重 |
缺陷组件
组件名称 | 版本 | 最小修复版本 | 依赖关系 | 修复建议 |
---|---|---|---|---|
torch | 1.12 | 1.13.1 | 间接依赖 | 建议修复 |
setuptools | 39.2.0 | 65.5.1 | 间接依赖 | 可选修复 |
numpy | 1.24.1 | 间接依赖 | 可选修复 | |
gradio | 3.34.0 | 间接依赖 | 可选修复 |
许可证风险
许可证类型 | 相关组件 | 许可证风险 |
---|---|---|
MIT | 14 | 低 |
Apache-2.0 | 8 | 低 |
自定义许可证 | 8 | 低 |
BSD-3-Clause | 4 | 低 |
BSD-2-Clause | 2 | 低 |
SBOM清单
组件名称 | 组件版本 | 是否直接依赖 | 仓库 |
---|---|---|---|
ColoTensor | 间接依赖 | pip | |
_DistSpec | 间接依赖 | pip | |
ConvModuleHandler | 间接依赖 | pip | |
imsave | 间接依赖 | pip | |
DistSpecManager | 间接依赖 | pip | |
LlamaTokenizer | 间接依赖 | pip | |
llama_load_quant | 间接依赖 | pip | |
degradation_fn_bsr_light | 间接依赖 | pip | |
BLOOMRM | 间接依赖 | pip | |
DAG_MLP | 间接依赖 | pip | |
ProfilerActivity | 间接依赖 | pip | |
opencv-python | 4.6.0.66 | 间接依赖 | pip |
replace_xformers | 间接依赖 | pip | |
pg_parse_args | 间接依赖 | pip | |
contextmanager | 间接依赖 | pip | |
broadcast | 间接依赖 | pip | |
Tuple | 间接依赖 | pip | |
DummyDataloader | 间接依赖 | pip | |
register_leaf_module_impl | 间接依赖 | pip | |
trange | 间接依赖 | pip | |
RpcTestModel | 间接依赖 | pip | |
BloomTokenizerFast | 间接依赖 | pip | |
DynamicGradScaler | 间接依赖 | pip | |
Iterable | 间接依赖 | pip | |
symbolic_trace | 间接依赖 | pip | |
get_gpt2_components | 间接依赖 | pip | |
HostInfoList | 间接依赖 | pip | |
BertForPreTraining | 间接依赖 | pip | |
imread | 间接依赖 | pip | |
hooks | 间接依赖 | pip | |
AutoConfig | 间接依赖 | pip | |
Iterator | 间接依赖 | pip | |
get_batch_size | 间接依赖 | pip | |
open-clip-torch | 2.7.0 | 间接依赖 | pip |
spawn | 间接依赖 | pip | |
bias_addition_method | 间接依赖 | pip | |
Graph | 间接依赖 | pip | |
MemoryEfficientAttentionCutlassOp | 间接依赖 | pip | |
reduce_by_batch_2d | 间接依赖 | pip | |
nn | 间接依赖 | pip | |
get_tensorboard_writer | 间接依赖 | pip | |
tensorboard | 2.14.0 | 间接依赖 | pip |
requirements.txt | 间接依赖 | pip | |
load_moe_model | 间接依赖 | pip | |
Linear | 间接依赖 | pip | |
rpc_is_initialized | 间接依赖 | pip | |
CollectiveCommPattern | 间接依赖 | pip | |
MemoryCost | 间接依赖 | pip | |
Generator | 间接依赖 | pip | |
_unflatten_dense_tensors | 间接依赖 | pip | |
flash_attn_func | 间接依赖 | pip | |
ColoGraphModule | 间接依赖 | pip | |
tqdm | 4.61.2 | 间接依赖 | pip |
ShardConfig | 间接依赖 | pip | |
SHARD_COST | 间接依赖 | pip | |
Repository | 间接依赖 | pip | |
numpy | 1.24.1 | 间接依赖 | pip |
imageio | 2.9.0 | 间接依赖 | pip |
run_on_environment_flag | 间接依赖 | pip | |
black | 23.9.1 | 间接依赖 | pip |
check_AB | 间接依赖 | pip | |
Repad | 间接依赖 | pip | |
OUTPUT_SAVED_MOD | 间接依赖 | pip | |
ColoTensorSpec | 间接依赖 | pip | |
List | 间接依赖 | pip | |
fastapi | 0.85.1 | 间接依赖 | pip |
clear_cache_before_run | 间接依赖 | pip | |
DDPMScheduler | 间接依赖 | pip | |
six | 1.16.0 | 间接依赖 | pip |
ColoTracer | 间接依赖 | pip | |
omegaconf | 2.1.1 | 间接依赖 | pip |
Dialogue | 间接依赖 | pip | |
ShardingNotDivisibleError | 间接依赖 | pip | |
ConvFunctionHandler | 间接依赖 | pip | |
BertDualHead | 间接依赖 | pip | |
pudb | 2019.2 | 间接依赖 | pip |
Policy | 间接依赖 | pip | |
DEPTH | 间接依赖 | pip | |
ComputeSpec | 间接依赖 | pip | |
Experts | 间接依赖 | pip | |
https | 间接依赖 | pip | |
BLOOMActor | 间接依赖 | pip | |
uniform_split_pass | 间接依赖 | pip | |
TensorState | 间接依赖 | pip | |
AutoTokenizer | 间接依赖 | pip | |
pydantic | 1.10.2 | 间接依赖 | pip |
UNetModel | 间接依赖 | pip | |
pandas | 1.4.1 | 间接依赖 | pip |
CLIPTokenizer | 间接依赖 | pip | |
FrozenCLIPTextEmbedder | 间接依赖 | pip | |
bias_dropout_add_fused_train | 间接依赖 | pip | |
Lars | 间接依赖 | pip | |
ModulePolicyDescription | 间接依赖 | pip | |
register_leaf_module | 间接依赖 | pip | |
parse_args | 间接依赖 | pip | |
gradio | 3.34.0 | 间接依赖 | pip |
ComputePattern | 间接依赖 | pip | |
GPTCritic | 间接依赖 | pip | |
Node | 间接依赖 | pip | |
LearningRateMonitor | 间接依赖 | pip | |
CLIPTextModel | 间接依赖 | pip | |
assert_equal_in_group | 间接依赖 | pip | |
FeatureFusionBlock_custom | 间接依赖 | pip | |
_make_pretrained_vitb_rn50_384 | 间接依赖 | pip | |
TensorMetadata | 间接依赖 | pip | |
Matmul_ABT_2p5D | 间接依赖 | pip | |
PromptDataset | 间接依赖 | pip | |
ProcessGroup | 间接依赖 | pip | |
model | 间接依赖 | pip | |
autoflake | 2.2.1 | 间接依赖 | pip |
colo_model_data_tensor_move | 间接依赖 | pip | |
is_compatible_with_meta | 间接依赖 | pip | |
MLP | 间接依赖 | pip | |
Lamb | 间接依赖 | pip | |
Embedding | 间接依赖 | pip | |
HiddenParallelEmbedding | 间接依赖 | pip | |
torch | 1.12 | 间接依赖 | pip |
CPUAdam | 间接依赖 | pip | |
ListConfig | 间接依赖 | pip | |
SFTDataset | 间接依赖 | pip | |
get_sam_flash_attention_forward | 间接依赖 | pip | |
flash_attn | 2.0.5 | 间接依赖 | pip |
HiddenParallelGPTLMHead1D | 间接依赖 | pip | |
setuptools | 39.2.0 | 间接依赖 | pip |
ViTConfig | 间接依赖 | pip | |
assert_close_loose | 间接依赖 | pip | |
solver_rotor | 间接依赖 | pip | |
CommSpec | 间接依赖 | pip | |
model_zoo | 间接依赖 | pip | |
dtype | 间接依赖 | pip | |
SUPPORT_LAZY | 间接依赖 | pip | |
bias_addition_function | 间接依赖 | pip | |
ListCache | 间接依赖 | pip | |
Callback | 间接依赖 | pip | |
get_lr_scheduler | 间接依赖 | pip | |
ImageDraw | 间接依赖 | pip | |
_create_vision_transformer | 间接依赖 | pip | |
OperationDataType | 间接依赖 | pip | |
split_batch_2p5d | 间接依赖 | pip | |
find_packages | 间接依赖 | pip | |
OPTForCausalLM | 间接依赖 | pip | |
_DistCommMethod | 间接依赖 | pip | |
MoeLayer | 间接依赖 | pip | |
TPExperts | 间接依赖 | pip | |
CostGraph | 间接依赖 | pip | |
BertLayer | 间接依赖 | pip | |
TransposeHandler | 间接依赖 | pip | |
balanced_split_pass | 间接依赖 | pip | |
check_state_dict_equal | 间接依赖 | pip | |
get_model | 间接依赖 | pip | |
locust | 2.11.0 | 间接依赖 | pip |
check_lazy_init | 间接依赖 | pip | |
convert_to_apex_amp | 间接依赖 | pip | |
protobuf | 3.20.0 | 间接依赖 | pip |
AutoencoderKL | 间接依赖 | pip | |
streamlit | 1.11.1 | 间接依赖 | pip |
LinearModuleHandler | 间接依赖 | pip | |
OUTPUT_SAVED_OPS | 间接依赖 | pip | |
DuplicatedShardingDimensionError | 间接依赖 | pip | |
PermuteHandler | 间接依赖 | pip | |
GPTActor | 间接依赖 | pip | |
SentencePiece | 0.1.99 | 间接依赖 | pip |
ChatGLMModel | 间接依赖 | pip | |
Trainer | 间接依赖 | pip | |
distribute_tensor | 间接依赖 | pip | |
sentencepiece | 0.1.99 | 间接依赖 | pip |
packaging | 23.1 | 间接依赖 | pip |
custom_bwd | 间接依赖 | pip | |
webdataset | 0.2.5 | 间接依赖 | pip |
ColCaiQuantLinear | 间接依赖 | pip | |
low_resource_init | 间接依赖 | pip | |
Matmul_AB_2p5D | 间接依赖 | pip | |
fbgemm-gpu | 0.2.0 | 间接依赖 | pip |
profile | 间接依赖 | pip | |
get_leaf | 间接依赖 | pip | |
ShardingSpec | 间接依赖 | pip | |
make_experience_batch | 间接依赖 | pip | |
titans | 0.0.7 | 间接依赖 | pip |
Any | 间接依赖 | pip | |
INPUT_GROUP_3D | 间接依赖 | pip | |
cast_tensor_to_fp16 | 间接依赖 | pip | |
colossalai | 0.3.2 | 间接依赖 | pip |
LowLevelZeroStrategy | 间接依赖 | pip | |
ShardingStrategy | 间接依赖 | pip | |
linearize | 间接依赖 | pip | |
SeqLenInfo | 间接依赖 | pip | |
ViTForImageClassification | 间接依赖 | pip | |
jdump | 间接依赖 | pip | |
gpipe_dp_split_pass | 间接依赖 | pip | |
uvicorn | 0.19.0 | 间接依赖 | pip |
PreTrainedTokenizer | 间接依赖 | pip | |
Callable | 间接依赖 | pip | |
LinearFunctionHandler | 间接依赖 | pip | |
recv_backward | 间接依赖 | pip | |
jload | 间接依赖 | pip | |
NOT_NVML | 间接依赖 | pip | |
Union | 间接依赖 | pip | |
BertConfig | 间接依赖 | pip | |
compatibility | 间接依赖 | pip | |
DropoutForReplicatedInput | 间接依赖 | pip | |
flash_attn_varlen_func | 间接依赖 | pip | |
bias_dropout_add_fused_inference | 间接依赖 | pip | |
custom_fwd | 间接依赖 | pip | |
get_timers | 间接依赖 | pip | |
prepare_buffers | 间接依赖 | pip | |
ChatGLMForConditionalGeneration | 间接依赖 | pip | |
albumentations | 1.3.0 | 间接依赖 | pip |
triton | 2.0.0.dev20221202 | 间接依赖 | pip |
chen_greedy | 间接依赖 | pip | |
MetaInfoProp | 间接依赖 | pip | |
parameterize | 间接依赖 | pip | |
ParallelMode | 间接依赖 | pip | |
DropoutForParallelInput | 间接依赖 | pip | |
degradation_fn_bsr | 间接依赖 | pip | |
GPTLMLoss | 间接依赖 | pip | |
DictConfig | 间接依赖 | pip | |
setup | 间接依赖 | pip | |
_AliasATen | 间接依赖 | pip | |
all_reduce | 间接依赖 | pip | |
check_grad | 间接依赖 | pip | |
assign_bfs_level_to_nodes | 间接依赖 | pip | |
GeminiStrategy | 间接依赖 | pip | |
get_data | 间接依赖 | pip | |
RingAV | 间接依赖 | pip | |
split_with_split_nodes_pass | 间接依赖 | pip | |
GeminiPlugin | 间接依赖 | pip | |
BloomForCausalLM | 间接依赖 | pip | |
get_tokenizer | 间接依赖 | pip | |
diffusers | 0.5.0 | 间接依赖 | pip |
build_model | 间接依赖 | pip | |
OperationData | 间接依赖 | pip | |
einops | 0.3.0 | 间接依赖 | pip |
convert_to_naive_amp | 间接依赖 | pip | |
OPTActor | 间接依赖 | pip | |
LlamaForCausalLM | 间接依赖 | pip | |
tmm_models | 间接依赖 | pip | |
SolverOptions | 间接依赖 | pip | |
DataloaderOption | 间接依赖 | pip | |
seed | 间接依赖 | pip | |
EncoderUNetModel | 间接依赖 | pip | |
sanic | 22.9.0 | 间接依赖 | pip |
recv_forward | 间接依赖 | pip | |
memory_efficient_attention | 间接依赖 | pip | |
LowLevelZeroPlugin | 间接依赖 | pip | |
SequenceParallelDataIterator | 间接依赖 | pip | |
save_moe_model | 间接依赖 | pip | |
matmul | 间接依赖 | pip | |
ColoParameter | 间接依赖 | pip | |
CaiQuantLinear | 间接依赖 | pip | |
palm_pytorch | 间接依赖 | pip | |
DataLoader | 间接依赖 | pip | |
convert_to_torch_amp | 间接依赖 | pip | |
GPT2LMHeadModel | 间接依赖 | pip | |
ConstantGradScaler | 间接依赖 | pip | |
Dict | 间接依赖 | pip | |
StatefulTensor | 间接依赖 | pip | |
NetflixDataset | 间接依赖 | pip | |
SupervisedDataset | 间接依赖 | pip | |
Set | 间接依赖 | pip | |
Extension | 间接依赖 | pip | |
nullcontext | 间接依赖 | pip | |
HostInfo | 间接依赖 | pip | |
_send_object | 间接依赖 | pip | |
Partition | 间接依赖 | pip | |
ABC | 间接依赖 | pip | |
Sequence | 间接依赖 | pip | |
RingQK | 间接依赖 | pip | |
_flatten_dense_tensors | 间接依赖 | pip | |
set_tuning_params | 间接依赖 | pip | |
PartitionInputVal | 间接依赖 | pip | |
get_profile_context | 间接依赖 | pip | |
Enum | 间接依赖 | pip | |
colo_set_process_memory_fraction | 间接依赖 | pip | |
check_ABT | 间接依赖 | pip | |
SUPPORT_FLASH | 间接依赖 | pip | |
check_topo | 间接依赖 | pip | |
_pair | 间接依赖 | pip | |
Critic | 间接依赖 | pip | |
HybridAdam | 间接依赖 | pip | |
_cfg | 间接依赖 | pip | |
rerun_if_address_is_in_use | 间接依赖 | pip | |
BATCH_SIZE | 间接依赖 | pip | |
BufferItem | 间接依赖 | pip | |
_make_pretrained_vitb16_384 | 间接依赖 | pip | |
WEIGHT_GROUP_3D | 间接依赖 | pip | |
sleep | 间接依赖 | pip | |
Image | 间接依赖 | pip | |
Actor | 间接依赖 | pip | |
ldm | 间接依赖 | pip | |
FrozenClipImageEmbedder | 间接依赖 | pip | |
ALLGATHER_COST | 间接依赖 | pip | |
HfFolder | 间接依赖 | pip | |
test-tube | 0.7.5 | 间接依赖 | pip |
torchrec | 0.2.0 | 间接依赖 | pip |
split_batch_2d | 间接依赖 | pip | |
BertLMHeadModel | 间接依赖 | pip | |
GPT2Config | 间接依赖 | pip | |
FFNExperts | 间接依赖 | pip | |
netflix_collator | 间接依赖 | pip | |
ChatPromptProcessor | 间接依赖 | pip | |
imageio-ffmpeg | 0.4.2 | 间接依赖 | pip |
BLOOMCritic | 间接依赖 | pip | |
transformers | 4.33.0 | 间接依赖 | pip |
TorchDDPPlugin | 间接依赖 | pip | |
get_micro_batch | 间接依赖 | pip | |
ShapeConsistencyManager | 间接依赖 | pip | |
auto | 间接依赖 | pip | |
sanic_ext | 22.9.0 | 间接依赖 | pip |
assert_close | 间接依赖 | pip | |
Solver | 间接依赖 | pip | |
colo_model_data_tensor_move_inline | 间接依赖 | pip | |
ninja | 1.11.1 | 间接依赖 | pip |
HhRlhfDataset | 间接依赖 | pip | |
RmStaticDataset | 间接依赖 | pip | |
DistPlacementPattern | 间接依赖 | pip | |
ShardFormer | 间接依赖 | pip | |
Interpolate | 间接依赖 | pip | |
DDPStrategy | 间接依赖 | pip | |
get_batch_for_sequence_parallel | 间接依赖 | pip | |
get_args | 间接依赖 | pip | |
abstractmethod | 间接依赖 | pip | |
coverage | 7.2.3 | 间接依赖 | pip |
lightning | 1.9.0 | 间接依赖 | pip |
torchmetrics | 0.7 | 间接依赖 | pip |
ModelAttribute | 间接依赖 | pip | |
SolverFactory | 间接依赖 | pip | |
MissCacheError | 间接依赖 | pip | |
forward_fn | 间接依赖 | pip | |
_single | 间接依赖 | pip | |
reduce_by_batch_2p5d | 间接依赖 | pip | |
OPTCritic | 间接依赖 | pip | |
all_gather | 间接依赖 | pip | |
tm_models | 间接依赖 | pip | |
colo_device_memory_capacity | 间接依赖 | pip | |
cast_tensor_to_bf16 | 间接依赖 | pip | |
Optional | 间接依赖 | pip | |
_recv_object | 间接依赖 | pip | |
Dataset | 间接依赖 | pip |